Model-free inverse reinforcement learning with multi-intention, unlabeled, and overlapping demonstrations.
Ariyan BighashdelPavol JancuraGijs DubbelmanPublished in: Mach. Learn. (2023)
Keyphrases
- model free
- inverse reinforcement learning
- temporal difference
- reinforcement learning
- reinforcement learning algorithms
- function approximation
- policy iteration
- preference elicitation
- active learning
- unsupervised learning
- reward function
- semi supervised learning
- monte carlo
- unlabeled data
- neural network
- labeled data
- training data
- decision makers
- prior knowledge
- training set
- feature space
- learning algorithm